app.py

import gradio as gr
from PIL import Image
import google.generativeai as genai
from IPython.display import Markdown

api_key = 'AIzaSyB7-RzBwTAfVA-7ZGk2mEOQwOxshpwzhpM' # put your API key here
genai.configure(api_key=api_key)
geminiModel = genai.GenerativeModel(model_name='gemini-pro-vision')
title="DocTR OCR (PDL Demo)"
description="Upload an image to get the OCR results !"

def to_markdown(text):
  text = text.replace('•', '  *')
  return Markdown(textwrap.indent(text, '> ', predicate=lambda _: True))
    
def greet(img):
    # generate text
    prompt = "the input text in vietnamese, please add accend and take this peace of information and give all the information in point wise better format also give some recomendation related to them: " + res
    # print(prompt)
    response  = geminiModel.generate_content([prompt,img], stream=True)
    response.resolve()
    print(response)
    res = response.text
    _output_name = "RESULT_OCR.txt"
    _output_name_pdf="RESULT_OCR.pdf"

    open(_output_name, 'w').close() # clear file
    with open(_output_name, "w", encoding="utf-8", errors="ignore") as f:
        f.write(res)
        print("Writing into file")

    base64_encoded_pdfs = list()
    for i, (xml, img) in enumerate(zip(xml_outputs, doc)):
      xml_element_tree = xml[1]
      parser.export_pdfa(_output_name_pdf,
            hocr=xml_element_tree, image=img)
      with open(_output_name_pdf, 'rb') as f:
            base64_encoded_pdfs.append(base64.b64encode(f.read()))
    return res, _output_name, _output_name_pdf

demo = gr.Interface(fn=greet,
                    inputs=gr.Image(type="pil"),
                    outputs=["text", "file","file"],
                    title=title,
                    description=description
                    )

demo.launch(debug=True)